A service control layer of a broadband metropolitan area network is a particular layer of a service access network connecting a core switched network. A device playing an important role is a BRAS, which mainly undertakes functions of aspects of service access control and user management.
With putting forward the high quality guarantee requirement of services provided by a network, it puts forward much higher requirement to the performance and the reliability of the BRAS product of the service control layer. Due to the functions undertook by the BRAS product and the location the BRAS locates in the network, it is demanded that the BRAS has the high reliability, so the concept of dual module hot spare of the BRAS emerges as the time require. The dual module hot spare of the BRAS is to deploy two BRAS products in the control layer of the network. During working, the two devices are set as that one is in a main state and another is in a standby state by negotiating or manually configuring, and the main device is used to control user accessing and record user information, and back up the user information on the backup device by real-time synchronization or batch synchronization. When the main device breaks down, it is able to realize main-standby switching rapidly. Since the user information has been already backed up to the standby state device, the user does not have the perception in the process of the main-standby switching of the devices, which ensures service quality and enhances user service experience.
The BRAS device airward connects to a core router (CR) device of the metropolitan area network, and notifies the network segment routing information of an address pool in which the user locates to the CR through a static route or an interior gateway protocol (IGP); and the CR does not maintain host routing information. The CR, after receiving the service flow, sends the service flow to the corresponding BRAS according to the network segment in which the destination address of the service flow locates. Between two BRASs as backup device of each other, the backup of the complete machine level, the line card level, the port level and the virtual sub-interface level can be supported, and a group of minimum abnormal switching units which back up each other are called a backup group. Among the main-standby devices, a plurality of backup groups can be established. For example, 20 ports of the main device perform backing up with the backup device one by one, there are 20 backup groups among the devices, and the backup group is identified through the port corresponding to each backup group.
The dual module hot spare mode of the BRAS supports switching of the single backup group, that is, when one port of the main device breaks down, the services at the backup group corresponding to that port is switched to the backup device; at the moment, it needs to change the routing information on the CR, so that the CR routes the service flow onto the backup device when the CR performs routing for the service flow of that backup group. When a plurality of backup groups share the same address pool, if a certain backup group performs the switching while other backup groups are not switched, then it is unable to change the routing information of the address pool on the CR, and the backup group which is switched already will be unable to receive the service flow normally. In order to prevent the problem that the CR is unable to route the service flow, a following method is adopted:
one address pool is configured individually for each backup group, for each backup group, the BRAS, as the main device, issues the routing information of the address pool to the CR; when the BRAS is abnormal and the switching is performed, the routing information of the main device of the backup group after switching is issued to the CR.
There is the following problem in above-mentioned process of the CR routing the service flow to the corresponding BRAS:
the address pool of each backup group cannot be shared with other backup groups, in this situation, a phenomenon may appear that the addresses in the address pool of one backup group has already been exhausted but there are a large number of residual addresses in the address pool of another backup group, which exists severe address waste, and affects the access service quality.